back to program ^

Key discussions:
- Review of key features in 4G/LTE-A for Interference Coordination and Coexistence.
- 5G NR features improving interference coordination and coexistence.
- Balancing Distributed vs Centralized Spectrum Sharing Coordination.
- Role of Measurement and Learning in Decision Making

back to program ^

After exploring the intense interest in Network Slicing, and some of the interesting claims being made for it, we will then discuss what is needed and what we have to meet real needs.

back to program ^

In the 5G-era, communication services and devices will be more diverse due to factors such as the diffusion of IoT devices. Network slicing has recently been attracted by network operators as one of solutions for adapting networks to such various types of services and devices. Network slicing is also expected to enable 3rd parties to instantiate and provide networks on demand, and it will lead to create new business chances on communication services.
Network slicing is a technology that enables to instantiate and handle separate virtual networks depending on requirements for each service, and would be realized by combination of various technologies, such as Network Function Virtualization (NFV), Software Defined Networking (SDN), and orchestration. Recently several SDOs are proceeding discussion of 5G including network slicing, and the requirements and architecture are being mostly firmed up.
Meanwhile, for enhancing applicability of network slices, some challenges are still remained. For example, providing appropriate communication quality from end to end on each slice would be required to guarantee required SLA, and from such aspect, resource management on whole networks across multiple administrative domains and isolation among slices on data plane would be important.
In this presentation, we will introduce some use cases and standard network technologies related to 5G and network slicing. In addition, we will show architecture for managing network resources depending on received service requirements from tenants, and requirements for data plane functionalities to appropriately establish/handle user traffic corresponding to assigned slices.

back to program ^

This presentation will provide performance comparisons between public DNS resolvers for both IPv4 and IPv6, based on measurements conducted from global locations over the course of 30 days. Trends, anomalies, the impact of China, and reasons for variation in performance will be explored.

back to program ^

Everything is being connected to the Cloud and Internet of Things, and network robots with big data analysis are creating important applications and services. The cloud network architecture is moving towards mega-cloud data centers (DCs) provided by companies such as Amazon and Google in combination with distributed small DCs or edge computers. While the traditional restrictions imposed by distance and bandwidth are being overcome by the development of advanced elastic optical interconnection that creates flexible slices, and high-bandwidth, modern applications impose more complex performance and quality of service requirements in terms of processing power, response time, and data amount. The rise in cloud performance must be matched by improvements in network performance.
Therefore, we propose an application-triggered cloud network architecture based on huge-bandwidth, logical local mesh, optical interconnections. This paper addresses edge/center cloud and edge/edge integration with the use of virtual machine (VM) migration. Edge computer interworking with center-cloud makes intelligent machine leaving platform for Autonomous Driving Vehicle (AD-car). Using the proposed architectures and technologies can realize energy-efficient and high-performance cloud service. In addition, demonstration system of AD-car control by edge computer are described. To grantee the response time of 10ms, VM migration techniques are apply to follow the vehicle movement. The flexible and programmable router based in resource pool architecture are used for this demonstration.
A number of Cloud-based applications are expected to offer advanced services to vehicles. Some of the applications will be running on VMs at Edge facilities vying for short network round trip latency with the vehicle. As the vehicle moves to other physical locations, the VM will be required to “Live migrate” along with its assigned vehicle, without interruption of service, in order to remain within the required of vehicle services, where horizontal here refers to the collection of the numerous Edge facilities next to the road infrastructure.
In this paper, we report demonstration results of dynamic VM migration using programmable edge node and optical interconnection under controlling with orchestrator.

back to program ^

Today's scientific computing applications and workflows operate on heterogeneous and vastly distributed infrastructures. Traditional human-in-the-loop service engineering approach met its insurmountable challenge in dealing with these very complex and diverse networked systems, including conventional and software defined networks, compute, storage, clouds and instruments. Orchestration is the key to integrate and coordinate the networked multi-services and automate end-to-end workflows. In this work, we present a model driven intelligent orchestration approach to this end-to-end automation, which is built upon a semantic modeling solution that supports the full stack of service integration, orchestration, abstraction, and intent and policy representation. We also present the design of a real-world orchestrator called StackV that is able to accommodate highly complex application scenarios such as Software Defined ScienceDMZ (SD-SDMZ) and Hybrid Cloud Inter-Networking (HCIN) by implementing this approach.

back to program ^

Northbound API has evolved for years. The first Northbound API started as a thin wrapper of OpenFlow protocol and developers should understand OpenFlow semantics to program network devices. This requires developers to think about how to program each network device and what instructions need to be installed into the device rather than network-wide policies. Open Network OS (ONOS) led by ONF as an open source project provides a Northbound API, called Intent Framework, which offers higher level abstraction than a thin OpenFlow wrapper. The speaker worked as a core developer of ONOS where he designed and developed Intent Framework which enables developers to specify a network-wide state that needs to be met through a simple API. The speaker will talk about what he considered when designing Intent Framework so that the API is easy-to-use for typical use cases and also extensible for developers who want to take full control of a network. Now, he is working on research and development of intent-based networking to improve the productivity of network operations. He will cover his recent activities and thoughts towards intent-based networking. Challenges for intent-based networking will also be expressed.

back to program ^

We present QoT (Quality of Transmission) management techniques of open source-based controller for open disaggregated optical transport networks. It would be difficult to guarantee and manage optical signal quality of transmission (QoT) in multi-vendor environment where is in the open architectures such as optical line system (OLS) and OpenROADM. Who should guarantee QoT and how QoT should be guaranteed are open for questions. We develop ONOS-based optical controller including a QoT management system which consists of QoT analysis and QoT monitoring functions. And then we demonstrate the QoT management system using an open disaggregated optical transport network testbed.

back to program ^

As disruptive technologies overhaul traditional industries, “liquid” workforces have emerged alongside the fast-growing gig economy in a phenomenon known as the Human Cloud. However, in spite of ever-accelerating innovation, why have companies been hesitant to integrate newly abundant digital tools into their internal and external processes? Often, slow traction in adopting contingent workers is related to lack of comfort, understanding, or trust in modern telecommunications tools. While companies and industries increasingly go extinct, firms and individuals who wish to survive will need to understand and participate in the Human Cloud or risk obsolescence. This talk will explain the business implications of disruptive technologies from the last decade, defining the implications of 5G applications, IoT, and others on the evolution of corporate and industrial Operational Models.

back to program ^

YANG has become the ubiquitous modelling language for configuring and operating network devices. Hundreds of YANG data models have been written over the last couple of years and published through various standards bodies and open source organisations.
Beyond simple configuration and operation of network devices, the coordination and use of these models has been limited. But, with the development of higher-level, service-oriented data models it is possible to consider a top-to-bottom YANG-based solution to service delivery.
This presentation will introduce some of the service-delivery models that allow network customers to describe the services they want to receive. These models express services such as L2VPN, L3VPN, and virtual networks as common parameters that allow standard and automated access.
We will then explain how these model can be combined with network resource models, network element models, and the protocol and device models to make a fully integrated YANG-based system that can be used to automatically deploy a variety of services on top of a multi-layer, multi-technology network. We will also show how the Abstraction and Control of Transport Networks (ACTN) architecture fits into this approach to manipulate network resources to build platforms on which to deliver services.

back to program ^

Communication networks are going to be composed of various slices of virtualized node and link resources. Network slices are configured with different amount of virtualized resources (e.g. CPU, memory, storage, and bandwidth), so that each slice can be able to optimally satisfy the quality of service (QoS) requirements of the service offered by the slice. Virtualized resources allocated to slices need to be dynamically adjusted to adapt the network capability to varying workload and network conditions. We present our ongoing research on mechanisms for the autonomic allocation and dynamic adjustment of resources to make the virtualized network system scalable, while optimally utilizing the resources. We present mechanisms for achieving three types of scalability: vertical scaling by adjusting virtualized resources allocated to a network function within a single node, horizontal scaling by adjusting the number of virtualized network nodes and servers, and internetwork scaling by the arbitration of limited resources of network nodes among many network slices that are configured in the node. We also discuss the preliminary performance evaluation results obtained from experiments.

back to program ^

I plan to discuss the policy and charging control for 5G, e.g. QoS control, gating control and charging control, it’s important for the operator and the user.

back to program ^

Segment Routing (SR) reintroduces source routing to networking be extending the core routing protocols. While SR has been defined for MPLS and IPv6 data planes, there are few problems with respect to increased path overhead in various deployments.
This presentation presents a new a new framework that is designed to overcome the SR challenges with a new routing paradigm, Preferred Path Routing (PPR). PPR is an innovative source routing architecture to signal explicit paths and per-hop processing including QoS from computation engines to network nodes leveraging distribution mechanisms of existing network routing protocols. PPR supports a wide range of existing forwarding planes including native IP (IPv4/IPv6) data planes. PPR also brings most needed optimizations to SR defined data planes (SR-MPLS, SRv6) in a fully-backward compatible fashion.
This presentation briefly describes the PPR architecture, its core features including optimized data planes, dynamic path QoS reservations to support high precision services. This also shows the applicability of the same in 5G transport networks and data center underlays.

back to program ^

Software-Defined Networking gave a trigger to introduce automated network operation into carrier networks. Recently, closed-loop network operation is proposed to improve availability and agility for carrier networks.
In this talk, we present intent-based networking approach to realize closed-loop operation. We have developed automated workflow designer that can re-program workflow and output configuration to change network to desired state. We are also developing network system designer that can design network functions (include servers and virtual machines), system configuration and parameters (IP/VLAN, etc.). Those technologies enable not only automated network operation but autonomous network design and operation.

back to program ^

Historically, IP addresses have provided information as to both the location in the network of communicating parties, and the identity of those parties. This tutorial will discuss some of the many problems caused by combining these, and then review approaches to solving these problems, with a focus on the ILNP, LISP, and ILA approaches.

back to tutorial ^

Our lifestyle will continue to push the technology boundaries even in post 5G era. When we try to imagine future life, a few things come to mind immediately, (a) that there will be a lot more automation; as machines perform tasks on our behalf, information will be collected and generated from multiple sources, creating humungous amounts of data, (b) more things will operate at system level, not in isolation; and will demand even tighter delivery guarantees than 5G at the same time coordinating distributed intelligence all over the connectivity fabric, (c) the lines between real and digital will get even fuzzier. We will advance from virtual, augmented, and mixed, immersive realities to a holographically rendered world in which communication networks will make distances immaterial.
In this talk we will describe the limitations of the current Internet architecture using some challenging and forward-looking scenarios such as holographic type communications, extremely fast response in critical situations and high-precision communication demands of emerging market verticals. These future use cases that our society would like to see and will benefit from, are not necessarily possible in the realm of current incumbent infrastructures. We also describe what kinds of network architecture and the enabling mechanisms and technologies that are suitable for such novel scenarios. For example, we will describe a new IP protocol suite that combines best effort and high-precision. It will deliver on the demands of the next wave of networking applications.
The talk includes a brief description of the Network 2030 initiative. We define the vision of Network 2030 as a progression of the capabilities of network infrastructures at the present and in IMT-2020, which are then extrapolated into the needs of future applications circa 2030. Network 2030 capabilities allow networks to aim to implement innovative networking algorithms that can be used for the novel schemes that will vastly improve network performance, security, mobility, and overall user experience. Network 2030 should make existing services better.
However, it is a huge job to bring the Internet to the truly next generation, and requires multiple companies, universities, SDOs to work together. Network 2030 means nothing without cooperation, collaboration, buy in from service providers, vendors and application developers. Without proper standardization and common vision, Network 2030 will not be possible. To this end ITU-T SG13 has created a new Focus Group on Network 2030 (FG NET-2030) which intends to study and advance international networking technologies, will investigate the future network architecture, requirements, use cases, and capabilities of the networks for the year 2030 and beyond.

back to tutorial ^